Commit History

v2 판독기 문서화 — 미지 모델에서 0.7752 (JEV 0.7521), 우리 표에서는 반대인 것도 명시
756e423
verified

SeaWolf-AI commited on

v2 곡선 판독기 추가 — 여러 모델의 답을 섞어 적합. 처음 보는 모델에서 0.7752 (JEV 0.7521)
0d157d1
verified

SeaWolf-AI commited on

정정: 폐기값(0.7364/0.7282) → 정본(0.7272/0.7289) + 정정 고지 추가
6cbd8ac
verified

SeaWolf-AI commited on

card: limitation — score shifts when the answering-model mixture changes (measured)
6f9b52f
verified

SeaWolf-AI commited on

card: generate candidates and let the verifier pick (+2.31 pp); majority voting measured as a loss
0c05b23
verified

SeaWolf-AI commited on

card: JEV-compatible API, calibration, and what to do on a low score
faf48c7
verified

SeaWolf-AI commited on

config: corrected per-domain LODO (curved probe), leaderboard 0.7282, baseline 0.6223
7491b79
verified

SeaWolf-AI commited on

docs: cross-link ZTC ecosystem badges
6338a5c
verified

SeaWolf-AI commited on

docs: cross-link ZTC ecosystem badges
6fef719
verified

SeaWolf-AI commited on

card: leaderboard measurements (0.7282), per-domain vs baseline, two probes documented
b6974a8
verified

SeaWolf-AI commited on

add runnable example for the curved probe
8617449
verified

SeaWolf-AI commited on

add curved probe (256 anchors, 5.3MB) - reproduces the leaderboard AUC 0.7282
0f453c5
verified

SeaWolf-AI commited on

card: full English rewrite — LODO protocol, per-domain table, baselines, limits
bd3c52e
verified

SeaWolf-AI commited on

ZTC-Judge-27B v1.0 — 자립형 (가중치+프로브+추론코드)
24b4119
verified

SeaWolf-AI commited on

ZTC-Judge-27B probe — 5 domains, 2018 items, cv_auc 0.7999 z=25.68
e238b68
verified

SeaWolf-AI commited on

initial commit
1a17f30
verified

SeaWolf-AI commited on